Launched early 2026, the LCSF is a national initiative led by the Luxembourg House of Cybersecurity (LHC). It aims to democratise and industrialise the cybersecurity data economy; its Cybersecurity Data Space, the first European Data Space dedicated to cybersecurity, provides open data, including open data ready to use to train, validate or infer on AI models. "A Cybersecurity Data Space is a trusted, decentralised digital infrastructure that allows organisations and public bodies to share cybersecurity data and AI-ready datasets under a shared governance framework, while retaining full control and sovereignty over their information," explains Dr. Emilia Tantar, Head of LCSF.
The project is based on a threefold assessment. First, a need for data: "Complying with DORA, NIS2, the Cyber Resilience Act, the EU AI Act requires data and secure access to that data. A reason we are building this 'CyberCommons,'" Dr. Tantar says.
Second, economics: "Cybersecurity is now a closed economy. Data, software, hardware and standardisation live in proprietary circuits. Intelligence remains compartmentalised, difficult to reuse. And data is rarely prepared for AI," she regrets.
Finally, a geostrategic observation: "Many countries have understood their dependence on certain manufacturers," says Mika Lauhde, Cyber Commons Lead at the LCSF. "Open source answers questions such as strategic autonomy."
The situation requires therefore a paradigm shift: "We believe in moving from open source to trusted source, from data silos to shared data, and from de facto standards to open standards," Dr. Tantar adds.
"The Cybersecurity Data Space is a federated environment where organisations share cybersecurity data under common rules, on infrastructure laid by the LHC (comprising three data centres, access to MeluXina), under the LNDS guidance, as to align with European Common Data Spaces practice. Its members remain the owners of their data; each organisation retains control over who uses it and for what," explains Tantar. "Members gain cybersecurity capacity and autonomy, with data exchanged peer-to-peer".
The data covers threat intelligence, incidents, vulnerabilities, telemetry, and datasets for AI and research. The first iteration will focus on open data; the second will introduce an OTC market based on trust and interoperability. "Luxembourg does not work in isolation: we are part of the wider ecosystem of AI factories, and twelve other countries have committed to providing AI-ready data," says Tantar.
A closing workshop explored participants' needs and the key findings are:
· Shared data needs: Identify cross-sectoral data needs relevant to all companies, including data needed to prepare compliance with horizontal regulatory requirements. The needs reach further in the operational needs and include asset inventories, threat intelligence, incident response, quality frameworks, data sustainability and compliance, supported by a common vocabulary for interpreting risk data.
· Building trust and enabling data sharing: Establish clear anonymisation practices, accessible tools and funding, safeguards against misuse or retaliation, and transparency regarding data recipients—particularly to support SME participation.
· Fostering a trusted, informed community: Build a community where members know and trust one another. Strengthen awareness in the industrial sector to mitigate vendor lock-in risks, and provide training on cyber monitoring and available Luxembourgish and European tools.
· Securing organisational and legal commitment: Equip boards and management with the arguments and incentives to participate and share information, while clarifying the legal framework and boundaries for data sharing.